LS3 has pushed the boundaries in semantics-aware information retrieval and search—the cornerstones of Web search engines. One of the challenges for efficient Web search is the vocabulary mismatch problem, where the terms of the query do not match the terms of the document. Our research has focused on improving information retrieval processes by incorporating external resources such as knowledge graphs. The results include the efficient indexing of semantic information, the consideration of semantic relatedness for inverted index design, the use of semantic information during pseudo-relevance feedback, the design of relevance functions based on semantic similarity and semantics-aware table retrieval. We have also explored how neural embedding representations of both textual documents and semantic information from the knowledge graph can be used in retrieval and learning to rank. Our most recent research work on measuring query specificity for Web query difficulty prediction shows that it is possible to predict Web query difficulty for both ad hoc retrieval and question answering tasks by measuring structural characteristics of pre-trained neural embeddings without considering corpus-specific information.
Semantics-enabled Information Processing and Retrieval
Sample Publications
Lashkari, F., F. Ensan, Bagheri, and A. A. Ghorbani (May 2017). “Efficient indexing for semantic search”. Expert Systems with Applications, 73: 92–114, IF: 4.292.
Lashkari, F., Bagheri, and A. A. Ghorbani (May 2019). “Neural Embedding-based Indices for Semantic Search”. Information Processing and Management, 56(3): 733-755, IF: 3.892.
Keikha, A.*, F. Ensan, and Bagheri (June 2018). “Query Expansion Using Pseudo Relevance Feedback on Wikipedia”. Journal of Intelligent Information Systems, June 2018, 50 (3): 455-478, IF: 1.589.
Ensan, F. and Bagheri (2017). “Document Retrieval through Semantic Entity Linking”. In: The Tenth International Conference on Web Search and Data Mining (WSDM).
Bagheri, E. , F. Al-Obeidat (2020), “A Latent Model for Ad Hoc Table Retrieval”, In: 42nd European Conference on Information Retrieval (ECIR).
Bagheri, E., F. Ensan, and F. Al-Obeidat (July 2018). “Neural Word and Entity Embeddings for Ad hoc Retrieval”. Information Processing and Management, 54(2): 339–357, IF: 3.892.
Ensan, F., Bagheri, A. Zouaq, and A. Kouznetsov (2017). “An Empirical Study of Embedding Features in Learning to Rank”. In: The 26th ACM International Conference on Information and Knowledge Management (CIKM).
Arabzadeh, N.*, F. Zarrinkalam, Jovanovic, F. Al-Obeidat, E. Bagheri (July 2020), “Neural Embedding-based Specificity Metrics for Pre-Retrieval Query Performance Prediction”, Information Processing and Management, 57 (4): 102248, 2020, IF: 3.892.
Arabzadeh, N.*, F. Zarrinkalam, Jovanovic, F. Al-Obeidat, E. Bagheri (July 2020), “Neural Embedding-based Specificity Metrics for Pre-Retrieval Query Performance Prediction”, Information Processing and Management, 57 (4): 102248, 2020, IF: 3.892.